Though it has been more than 30 years since the first discovery of the classic Galactic black hole candidate (GBHC) Cygnus X-1, the X-ray emission mechanisms of GBHCs are still not well understood. It is generally agreed that black holes accrete materials from nearby objects or media to form accretion disks around them. The accretion disks can be heated up to a temperature above 1 keV. Such a hot disk emits X- and Gamma-rays, observations of which are essential for understanding the accreting and heating processes. Our multi-wavelength observations of GRS 1758-258 (a GBHC) in August 1997 revealed several properties that are important for constraining the emission model and the geometry of the accretion disk. (1) Its spectrum does not have a significant soft component. This implies that the cold optically thick disk must be small or be mostly covered by a optically thin hot corona. (2) The spectrum has an exponential cutoff around 200 keV, which means that the hot corona is largely thermal. (3) No significant iron lines are detected. This suggests that the reprocessing of the X-ray by the cold disk is negligible. We systematically analyzed RXTE archival data to study the energy dependency of the variability of X-ray flux from four hard X-ray sources: Cygnus X-1, GX 339-4, GRS 1758-258 and 1E 1740.7-2942. Cygnus X-1 was found to have flatter power density spectrum (PDS) shapes at higher energies, while the other three have energy independent PDS shapes. No current models can fully explain these results. A general trend was found among the four sources that the variability anti-correlates with the X-ray flux. We found that the 0.5--10 Hz quasi-periodic oscillations (QPOs) observed in GRS 1915+105 (a GBHC) has peculiar phase lag behaviors. When the QPO fundamental frequency is low (0.5--2.0 Hz), positive phase lags, which mean that hard photons arrive later than soft photons, were observed in both fundamental and first harmonic frequencies. The phase lags have opposite signs at the two frequencies when the fundamental frequency is high (2.0--4.5 Hz). Such strange behaviors can not be explained by current models. We found the flat radio spectrum observed in most GBHCs can be explained by a hybrid thermal/non-thermal plasma, which can also reproduce the power-law tail that have been observed in many GBHCs beyond the thermal X-ray spectrum.